摘要
为评估降水分布指数模型(DPEM)的精度问题,基于圣卡塔利娜岛4场次降雨7个观测站点的实测数据,分别运用距离倒数加权插值模型(IDW)、修正距离倒数插值模型(RIDM)、空间线性插值模型(SLIM),通过交叉验证,比较了四种模型的插值精度。结果表明,对不同验证站进行插值时,各方法呈现不同特征。验证站位于样本地理分布边缘时,SLIM、IDW模拟误差有增大趋势;RIDM考虑高程作为权重因子后,模拟结果优于IDW;而DPEM模拟产生的误差更小,且能够反映降水中心的趋势位置。该评估比较对降水分布指数模型在降水数据输入时有指导意义。
To evaluate the performance of distributed precipitation exponential model （DPEM）, we compared the ac- curacies of four spatial interpolation methods including IDW, RIDM, SLIM and DPEM by cross validation. The precipi- tation data are from seven observatories on the Santa Catalina Islam The results show that each method has different trend when the validated station is changing. SLIM and IDW incline to produce more simulation error when the validated station is in the edge of geographic distribution of the sample. RIDM is more accurate than IDW for considering the eleva- tion to be a part of the weighting factor. Compared with three methods above, the simulation error of DPEM is much lower and it could reflect the approximate center of the daily precipitation. This evaluation has guiding meaning for the in- put of the precipitation data in distributed hydrological model.
出处
《水电能源科学》
北大核心
2015年第6期19-21,30,共4页
Water Resources and Power
基金
国家自然科学基金项目(41471025
40971021)
山东省自然科学基金项目(ZR2014DM004)
关键词
降水分布指数模型
空间插值模型
降水中心
交叉验证
distributed precipitation exponential model
spatial interpolation model
precipitation centre
cross validation
作者简介
张升堂（1970-），男，副教授，研究方向为水文水资源，E-mail：zst0077@163．com
